ggplot2

Example data

head(mpg)

## # A tibble: 6 × 11
##   manufacturer model displ  year   cyl      trans   drv   cty   hwy    fl
##          <chr> <chr> <dbl> <int> <int>      <chr> <chr> <int> <int> <chr>
## 1         audi    a4   1.8  1999     4   auto(l5)     f    18    29     p
## 2         audi    a4   1.8  1999     4 manual(m5)     f    21    29     p
## 3         audi    a4   2.0  2008     4 manual(m6)     f    20    31     p
## 4         audi    a4   2.0  2008     4   auto(av)     f    21    30     p
## 5         audi    a4   2.8  1999     6   auto(l5)     f    16    26     p
## 6         audi    a4   2.8  1999     6 manual(m5)     f    18    26     p
## # ... with 1 more variables: class <chr>

Quickstart

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point()

A plot has 3 main components:

• data
• a set of aesthetics
• a set of layers (mainly geometries)

ggplot(mpg,   ## data
       aes(x = displ, y = hwy)) +  ## aesthetics
    geom_point() ## layer

add geoms

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_line()

aesthetics

• x
• y
• z
• color
• fill
• size
• text
• shape
• alpha
• group

Set color aesthetic to manufacturer variable

ggplot(mpg, aes(x = displ, y = hwy, color = manufacturer)) + geom_point()

Set size aesthetic

ggplot(mpg, aes(x = displ, y = hwy, size = cyl)) + geom_point()

Alpha aesthetic

ggplot(mpg, aes(x = displ, y = hwy, size = cyl, color = class)) + geom_point(alpha = 0.5)

Set an aesthetic to a fixed value

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point(aes(color = "blue"))

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point(color = "blue")

Set an aestetic to a discrete numeric value

ggplot(mpg, aes(x = displ, y = hwy, color = cyl)) + geom_point() ## gradient scale

ggplot(mpg, aes(x = displ, y = hwy, color = factor(cyl))) + geom_point() ## discrete scale

Aesthetics can be defined inside geom

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point(aes(color = class))

Aesthetics defined inside ggplot apply to all geom

ggplot(mpg, aes(x = class, y = hwy, color = class)) + geom_boxplot() + geom_point()

a ggplot is an object

We can build a ggplot object in several steps

g <- ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() 
g ## or print(g)

g <- g + ggtitle("hwy vs displ")
g

geom

• one variable
  • continous
    • geom_density
    • geom_histogram
    • geom_freqpoly
  • discrete
    • geom_bar
• two variables
  • continous X, continuous Y
    • geom_point
    • geom_smooth
    • geom_text
  • discrete X, continuous Y
    • geom_bar
    • geom_boxplot
    • geom_violin
  • bivariate distribution
    • geom_density2d
    • geom_hex
    • geom_bin2d
  • continuous function
    • geom_line
    • geom_step
• three variables
  • • geom_contour
    • geom_tile

data distributions

For histograms and densities only x is mandatory

ggplot(mpg, aes(x = hwy)) + geom_histogram()

ggplot(mpg, aes(x = hwy)) + geom_density()

ggplot(mpg, aes(x = hwy)) + geom_freqpoly()

Changing bins

ggplot(mpg, aes(x = hwy)) + geom_histogram(bins = 30)

ggplot(mpg, aes(x = hwy)) + geom_histogram(binwidth = 2)

Manually setting breaks

ggplot(mpg, aes(x = hwy)) + geom_histogram(breaks = c(10, 15, 20, 25, 30, 40, 50))

Add fill or color

ggplot(mpg, aes(x = hwy)) + geom_density(fill = "blue", color = "red")

ggplot(mpg, aes(x = hwy)) + geom_density(aes(fill = class))

Boxplots, violins, jitter

boxplots need x and y aestetics

ggplot(mpg, aes(x = class, y = hwy)) + geom_boxplot()

ggplot(mpg, aes(x = class, y = hwy, fill = class)) + geom_boxplot()

Add another variable as fill or color will separate boxplots

ggplot(mpg, aes(x = class, y = hwy, fill = trans)) + geom_boxplot()

Violin plots

ggplot(mpg, aes(x = class, y = hwy)) + geom_violin()

Jitters

ggplot(mpg, aes(x = class, y = hwy)) + geom_jitter()

## control width and height of jitter
ggplot(mpg, aes(x = class, y = hwy)) + geom_jitter(width = 0.5, height = 0)

ggplot(mpg, aes(x = class, y = hwy)) + geom_boxplot() + geom_jitter(width = 0.5, height = 0)

Bars

ggplot(mpg, aes(x = manufacturer)) + geom_bar()

Adding color

ggplot(mpg, aes(x = manufacturer, fill = class)) + geom_bar()

Changing bars positions

ggplot(mpg, aes(x = manufacturer, fill = class)) + geom_bar(position = "stack")

ggplot(mpg, aes(x = manufacturer, fill = class)) + geom_bar(position = "dodge")

ggplot(mpg, aes(x = manufacturer, fill = class)) + geom_bar(position = "fill")

Bar plot from count data

d <- data.frame(group = c("a", "b", "c"), n = c(10, 25 , 14))

## error
## ggplot(d, aes(x = group, y = n)) + geom_bar()
ggplot(d, aes(x = group, y = n)) + geom_bar(stat = "identity")

add a smoother on the plot

Default smoother (loess if <1000 obs, gam otherwise)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_smooth()

Linear model

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_smooth(method = "lm")

Remove SE

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_smooth(method = "lm", se = FALSE)

Smooth by group

ggplot(mpg, aes(x = displ, y = hwy, color = class)) + geom_point() + geom_smooth(method = "lm", se = FALSE)

Order of layers in object construction

Order of layers matters

ggplot(mpg, aes(x = class, y = hwy)) + geom_jitter(width = 0.5, height = 0) + geom_boxplot() 

ggplot(mpg, aes(x = class, y = hwy)) + geom_boxplot() + geom_jitter(width = 0.5, height = 0) 

Scales

Scales are mainly controlled by scale_* functions. Their names are structured: *scale_‘aestetic name’_‘type’*

• scale_color_discrete: discrete scale for color aestetic
• scale_fill_gradient: continuous scale for fill aestetic

Discrete scales

ggplot(mpg, aes(x = displ, y = hwy, color = class)) + geom_point() + scale_color_discrete(h = c(20, 150))

ggplot(mpg, aes(x = displ, y = hwy, color = class)) + geom_point() + scale_color_brewer(palette = "Set2")

ggplot(mpg, aes(x = displ, y = hwy, color = factor(cyl))) + geom_point() + scale_color_manual(values = c("blue", "black", "red", "pink"))

ggplot(mpg, aes(x = displ, y = hwy, color = factor(cyl))) + geom_point() + scale_color_manual(values = c('8' = "blue", '6' = "black", '4' = "red", '5' = "pink"))

gradient scales

ggplot(mpg, aes(x = displ, y = hwy, color = hwy)) + geom_point()

## 2 colors gradient
ggplot(mpg, aes(x = displ, y = hwy, color = hwy)) + geom_point() + scale_color_gradient(low = "green", high = "red")

## diverging color gradient, define low, mid and high colors
ggplot(mpg, aes(x = displ, y = hwy, color = hwy)) + geom_point() + scale_color_gradient2(low = "green", high = "red", mid = "black", midpoint = 25)

## n colors gradient
ggplot(mpg, aes(x = displ, y = hwy, color = hwy)) + geom_point() + scale_color_gradientn(colours = terrain.colors(10))

xlim/ylim vs coord_cartesian

• xlim/ylim select data inside the range
• coord_cartesian zoom in the plot

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point()

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + xlim(c(2, 6))

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + coord_cartesian(xlim = c(2, 6))

Warning: when computing densities or smoothers, the plot will completely change by using xlim/ylim

ggplot(mpg, aes(x = displ)) + geom_density(fill = "gray50")

ggplot(mpg, aes(x = displ)) + geom_density(fill = "gray50") + xlim(c(2, 3))

ggplot(mpg, aes(x = displ)) + geom_density(fill = "gray50") + coord_cartesian(xlim = c(2, 3))

axis transformation

• flip axes

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + coord_flip()

• polar coordinates

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + coord_polar()

• reverse axis

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + scale_x_reverse()

• log axis

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + scale_x_log10()

• order of discrete scales

ggplot(mpg, aes(x = fl)) + geom_bar()

## change levels order 
mpg$fl2 <- factor(mpg$fl, levels = c("e", "p", "d", "r", "c"))
ggplot(mpg, aes(x = fl2)) + geom_bar()

## or use xlim
ggplot(mpg, aes(x = fl)) + geom_bar() + xlim(c("e", "p", "d", "r", "c"))

Themes

Themes allow to have a set of graphics with the same style. Some are predefined theme_default, theme_bw, them_minimal, theme_gray...

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + theme_bw()

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + theme_minimal()

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + theme_dark()

We can customize some elements in themes (see http://docs.ggplot2.org/dev/vignettes/themes.html for details)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + theme(panel.background = element_rect(fill = "wheat"), axis.text = element_text(size = 20))

Faceting

Facetting allows to display plot separately according to discrete variables

Two functions for facetting: - facet_wrap - facet_grid

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + facet_wrap(~class)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + facet_grid(.~class)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + facet_grid(class~drv)

Controling facet_wrap organisation

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + facet_wrap(~class, ncol = 2)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + facet_wrap(~class, scales = "free")

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + facet_wrap(~class, scales = "free_x")

All layers are faceted

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_smooth() + facet_grid(~drv)

add supplementary information on plot

add lines

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_hline(yintercept = 20)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_vline(xintercept = 4.5)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_abline(slope = -2, intercept = 35)

annotate

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + annotate(x = 4, y = 30, geom = "point", color = "red", size = 4)

ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + annotate(x = 4, y = 30, geom = "label", label = "annotation") 

add data from external data.frame

d_annot <- data.frame(xpos = c(2, 5, 8), ypos = c(20, 25, 40), lab = c("l1", "l2", "l3"))
ggplot(mpg, aes(x = displ, y = hwy)) + geom_point() + geom_label(aes(x = xpos, y = ypos, label = lab), data = d_annot)

Programming with ggplot2

ggplot2 use non standard evaluation (NSE). How to use variables in aes? aes_string allows standard evaluation.

my_var <- "displ"
ggplot(mpg, aes(x = my_var, y = hwy)) + geom_point()

ggplot(mpg, aes_string(x = my_var, y = "hwy")) + geom_point()

Interactivity

For interactive graphics see ggvis or plotly packages

require(ggvis)
ggvis(mpg, ~displ, ~hwy) %>% layer_points()

ggvis(mpg, ~displ, ~hwy) %>%  
    layer_smooths(span = input_slider(0.2, 1, value = 1), stroke := "blue") %>% 
    layer_points(size := input_slider(10, 200, value = 50))

require(plotly)
g <- ggplot(mpg, aes(x = displ, y = hwy)) + geom_point()
ggplotly(g)

Extensions

• ggrepel: smart geom_text and geom_label placement
• cowplot: several plots on one page
• ggraph: visualise graph
• ggmap: plot data on a map
• factoextra: visualise results of factorial analysis (PCA, CA, MCA, MFA, …)
• ggbio: visualise genomic data
• ggdendro: visualise dendrograms and trees
• ggthemes: additionnal themes

Ressources

• http://docs.ggplot2.org
• rstudio cheatsheet http://www.rstudio.com/wp-content/uploads/2015/04/ggplot2-cheatsheet.pdf
• ggplot2 book https://github.com/hadley/ggplot2-book